Effect of Matching Algorithm and Profile Shape on Pulsar Pulse Time of Arrival Uncertainties

For high-precision pulsar timing analysis and low-frequency gravitational wave detection, it is essential to accurately determine pulsar pulse times of arrival (ToAs) and associated uncertainties. To measure the ToAs and their uncertainties, various cross-correlation-based techniques can be employed. We develop methodologies to investigate the impact of the template-matching method, profile shape, signal-to-noise ratio of both template and observation on ToA uncertainties. These methodologies are then applied to data from the International Pulsar Timing Array. We demonstrate that the Fourier domain Markov chain Monte Carlo method is generally superior to other methods, while the Gaussian interpolation shift method outperforms other methods in certain cases, such as profiles with large duty cycles or smooth profiles without sharp features. However, it is important to note that our study focuses solely on ToA uncertainty, and the optimal method for determining both ToA and ToA uncertainty may differ.